| 融合Sentinel-2红边波段和Sentinel-1雷达波段影像的扎龙湿地信息提取 |
| |
| 引用本文： | 常文涛,王浩,宁晓刚,张翰超.融合Sentinel-2红边波段和Sentinel-1雷达波段影像的扎龙湿地信息提取[J].湿地科学,2020,18(1):10-19. |
| |
| 作者姓名： | 常文涛 王浩 宁晓刚 张翰超 |
| |
| 作者单位： | 山东科技大学测绘科学与工程学院,山东青岛266590;中国测绘科学研究院,北京100036;中国测绘科学研究院,北京100036 |
| |
| 基金项目： | 国家重点研发计划项目(2017YFC0404503);中国测绘科学研究院基本科研业务费专项项目(Y518002)资助。 |
| |
| 摘 要： | 以黑龙江流域中的扎龙湿地及其上游区域为研究区,将Sentinel-2红边波段和Sentinel-1雷达波段影像数据相结合,根据面向对象原理,采用随机森林算法,对研究区的湿地进行遥感分类和信息提取;利用3种特征变量集,进行实验对比,研究红边波段反射率和雷达后向散射系数对湿地信息提取的作用。研究结果表明,红边波段反射率和雷达后向散射系数对土地覆盖分类精度的提高起到了重要作用,两者结合得到的分类结果的总体精度达到了88.72%,Kappa系数为0.87,其中,水体、水田和沼泽的用户精度分别为100%、98.18%和91.37%。利用红边波段和雷达波段影像数据,分别使土地覆盖分类总体精度提高了5.26%和2.51%,红边波段影像数据对沼泽提取精度的提高贡献最大,使生产者精度提高了12.5%。
|
| 关 键 词： | 红边波段 雷达波段 Sentinel-1 Sentinel-2 面向对象 随机森林 |
Extraction of Zhalong Wetlands Information based on Images of Sentinel-2 Red-edge Bands and Sentinel-1 Radar Bands |
| |
| CHANG Wentao,WANG Hao,NING Xiaogang,ZHANG Hanchao.Extraction of Zhalong Wetlands Information based on Images of Sentinel-2 Red-edge Bands and Sentinel-1 Radar Bands[J].Wetland Science,2020,18(1):10-19. |
| |
| Authors: | CHANG Wentao WANG Hao NING Xiaogang ZHANG Hanchao |
| |
| Institution: | (College of Geomatics,Shandong University of Science and Technology,Qingdao 266590,Shandong,P.R.China;Chinese Academy of Surveying and Mapping,Beijing 100036,P.R.China) |
| |
| Abstract: | In this paper, the Zhalong Wetland and its upstream area in the Heilongjiang River Basin were selected as the study area, and the images of Sentinel-2 red-edge bands and Sentinel-1 radar bands were used combined to extract the information of the wetlands with a random forest algorithm method according to the object-oriented principle. Three sets of characteristic variables were used to study the effects of red edge band reflectance and radar backscatter coefficient on extraction of the wetland information. In this research, 3 types of characteristic variables were used for experimental comparison to study the effects of reflectance of rededge bands and radar backscattering coefficients on extraction of the wetland information. The results indicated that the reflectance of red-edge bands and radar backscattering coefficients played important roles in improving the accuracy of land cover classification, the overall accuracy of the classification result obtained by the combining the two kinds of image data reached 88.72%, and the Kappa coefficient was 0.87, while the user accuracy of the water body, paddy field and marsh showing 100%, 98.18% and 91.37% respectively. In addition, the use of images of red-edge bands and radar bands increased the overall accuracy of land cover classification by 5.26% and 2.51%, respectively. The application of images of red-edge bands produced the most improvement of extraction accuracy of the marshes, which increased the producer’s accuracy by 12.5%. |
| |
| Keywords: | red-edge bands radar bands Sentinel-1 Sentinel-2 object-oriented random forest |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
